vs
15:1
overall capability wins (1 ties across 17 scored). Gemini 2.5 Pro leads overall.
- Reasoning9:0
- Coding4:1
- Arena1:0
- Pricing2:1
- Speed1:1
- Specs1:0
Verdict
Gemini 2.5 Pro leads coding (widest gap: +34.4 on LiveCode); Gemini 2.5 Pro ranks higher on LMArena (+31 Elo); Gemini 2.5 Pro allows ~2.0x more max output (66K vs 33K).
- Gemini 2.5 Pro leads coding (widest gap: +34.4 on LiveCode)
- Gemini 2.5 Pro ranks higher on LMArena (+31 Elo)
- Gemini 2.5 Pro allows ~2.0x more max output (66K vs 33K)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Gemini 2.5 Pro
- $3.75
- GPT-4.1
- $4
Gemini 2.5 Pro is about 1.1x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
Family
Gemini
GPT-4 / o-series
License
Proprietary
Proprietary
Open weights
No
No
Release
Mar 25, 2025
Apr 14, 2025
Knowledge cutoff
-
2024-06
API / provider
OpenAI
Modalities
text, image, audio, video → text
text, image → text
Specs
Context window
1M
1M
Max output
66K
33K
Parameters
—
—
Pricing
Input $/1M
$1.25
$2
Output $/1M
$10
$8
Blended $/1M (3∶1)
$3.44
$3.50
Speed
tok/s
100
95
TTFT (s)
0.45
0.35
Reasoning
MMLU-Pro
86.2
80.6
GPQA Diamond
84.4
66.6
Humanity's Last Exam
22.5
4.2
AIME 2025
87.7
46.4
MATH-500
96.7
91.3
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
39.1
27.8
AA-LCR v1.1
69.0
68.3
CritPt
2.6
—
MMMU-Pro
74.9
61.2
IFBench
48.7
43.0
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
53.6
39.6
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
80.1
45.7
Terminal-Bench 2.1
28.5
—
Aider Polyglot
79.1
52.4
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
46.3
38.1
CursorBench
—
—
SWE-Rebench
23.5
30.1
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,226
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,446
1,415
| Metric | Gemini 2.5 Pro | GPT-4.1 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | — | |
| Family | Gemini | GPT-4 / o-series | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Mar 25, 2025 | Apr 14, 2025 | — |
| Knowledge cutoff | - | 2024-06 | — |
| API / provider | OpenAI | — | |
| Modalities | text, image, audio, video → text | text, image → text | — |
| Specs | |||
| Context window | 1M | 1M | tie |
| Max output | 66K | 33K | A +33K |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $1.25 | $2 | A +$0.75 |
| Output $/1M | $10 | $8 | B +$2 |
| Blended $/1M (3∶1) | $3.44 | $3.50 | A +$0.06 |
| Speed | |||
| tok/s | 100 | 95 | A +5.00 |
| TTFT (s) | 0.45 | 0.35 | B +0.10 |
| Reasoning | |||
| MMLU-Pro | 86.2 | 80.6 | A +5.6 pts |
| GPQA Diamond | 84.4 | 66.6 | A +17.8 pts |
| Humanity's Last Exam | 22.5 | 4.2 | A +18.3 pts |
| AIME 2025 | 87.7 | 46.4 | A +41.3 pts |
| MATH-500 | 96.7 | 91.3 | A +5.4 pts |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 39.1 | 27.8 | A +11.3 pts |
| AA-LCR v1.1 | 69.0 | 68.3 | A +0.7 pts |
| CritPt | 2.6 | — | — |
| MMMU-Pro | 74.9 | 61.2 | A +13.7 pts |
| IFBench | 48.7 | 43.0 | A +5.7 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 53.6 | 39.6 | A +14.0 pts |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 80.1 | 45.7 | A +34.4 pts |
| Terminal-Bench 2.1 | 28.5 | — | — |
| Aider Polyglot | 79.1 | 52.4 | A +26.7 pts |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 46.3 | 38.1 | A +8.2 pts |
| CursorBench | — | — | — |
| SWE-Rebench | 23.5 | 30.1 | B +6.6 pts |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,226 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,446 | 1,415 | A +31 Elo |
Gemini 2.5 Pro
- Context
- 1M / 66K out
- Parameters
- —
- Price
- $1.25 / $10
- Speed
- 100 tok/s · 0.45s TTFT
- Modalities
- text, image, audio, video → text
- License
- Proprietary
Previous-gen Google Pro model; still a capable long-context multimodal option at competitive rates.
GPT-4.1
- Context
- 1M / 33K out
- Parameters
- —
- Price
- $2 / $8
- Speed
- 95 tok/s · 0.35s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Previous-gen OpenAI instruction model with 1M context; still useful for long-document workloads.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.